I have cross posted looking at this as having three disciplines, acoustics,
electronics
and magnetics.
I would like to see ideas to design one with these objectives in mind.
Objectives:
Maximum audio from minimal (crystal radio) signal (high sensitivity)
High impedance, if we can get enough wire on the electromagnet, the matching
transformer could be eliminated. (100k to 1 meg ohms, tapped) That's a wish
:-)
Questions;
What should the diaphragm look like? low mass? with ridges? needs to be
magnetic at
some point.
Can we put an electromagnet on both sides in push pull? Can needs to be
sealed.
Earphone is connected with hollow tubing, what type of tube would be low
loss?
does length matter? I'm thinking stethoscope earphones till something better
is found.
Do we want the center of the diaphragm to move or the whole circumference to
breath?
How do we focus the magnetic field, if we do.
Thanks for your input, MikeK
I would point out that Electromagnetic phones (sound powered headphones)
are very commonly used with crystal radios and yes, that's why I want
this to be very high impedance.
So maybe we need 10.000 turns and high (AsubL?) core.
Thanks, MikeK
> Hi guys,
> I ran across this electromechanicoustic device (speaker) called a Gallows
> headphone.
> This is used as the speaker for a crystal radio.
> It is pictured about 2/3 of the way down the page.
> You will understand it's operation by its picture.
> http://www.hpfriedrichs.com/
>
I could not make head or tail out of these pictures. A well thought out
diagram would be much more helpful.
In general, you have a number of impedance matching problems. The
crystal set provides the source impedance. That power has to be matched
to a mechanical device like the stirrup in your ear. That in turn is
attached to a diaphragm which must be matched to the air. Everywhere in
this path, a mismatch will reduce the amount of power that can be
transferred. That is the trick.
Bill
--
An old man would be better off never having been born.
> In general, you have a number of impedance matching problems. The
> crystal set provides the source impedance. That power has to be matched
> to a mechanical device like the stirrup in your ear. That in turn is
> attached to a diaphragm which must be matched to the air. Everywhere in
> this path, a mismatch will reduce the amount of power that can be
> transferred. That is the trick.
>
> Bill
>
Yes matching is the trick, that's why I ask the experts.
MikeK
PS. a perfect match is never going to be made over the useful audio range,
hopefully just better than what's available now.
Impedance matching is only important when the objective is to maximize
power ransfer. Since the ear is only interested in pressure, maximizing
power transfer serves no useful purpose. In fact, The Designer actually
uses an impedance mismatch to create a 15-dB amplification of pressure at
the eardrum at about 3000Hz (quarter wave ear canal resonance). It is
this resonance combined with the resonance of the middle ear at about
1000Hz that provides the low threshold of audibility in humans that
exists between about 1000Hz and 3000Hz.
> Yes matching is the trick, that's why I ask the experts.
> MikeK
> PS. a perfect match is never going to be made over the useful audio range,
> hopefully just better than what's available now.
Indeed, that is the trick. I do not think that you will find an expert
that can make an almost perfect match across the band. You probably
won't find one who would try developing a good broadbamd match cheaply.
Good luck!
> Impedance matching is only important when the objective is to maximize
> power ransfer. Since the ear is only interested in pressure, maximizing
> power transfer serves no useful purpose. In fact, The Designer actually
> uses an impedance mismatch to create a 15-dB amplification of pressure at
> the eardrum at about 3000Hz (quarter wave ear canal resonance). It is
> this resonance combined with the resonance of the middle ear at about
> 1000Hz that provides the low threshold of audibility in humans that
> exists between about 1000Hz and 3000Hz.
Maximum power across the band is what you want. Exceptions would be if
you are listening to a a narrow audio band signal like tone modulation
of an rf signal detected by the crystal. The amplification will only
occur if the acoustic network is the equivalent of a lumped circuit.
Resonance at 3kHz is not going to help listen to an ordinary voice
signal.
> In article <27OdndD7C4IkNkzR...@giganews.com>,
> Answerman <answerman@home.> wrote:
>
>> Impedance matching is only important when the objective is to
>> maximize power ransfer. Since the ear is only interested in
>> pressure, maximizing power transfer serves no useful purpose. In
>> fact, The Designer actually uses an impedance mismatch to create a
>> 15-dB amplification of pressure at the eardrum at about 3000Hz
>> (quarter wave ear canal resonance). It is this resonance combined
>> with the resonance of the middle ear at about 1000Hz that provides
>> the low threshold of audibility in humans that exists between about
>> 1000Hz and 3000Hz.
>
> Maximum power across the band is what you want.
No, maximum pressure at the eardrum across the band is what you want.
> Exceptions would be if
> you are listening to a a narrow audio band signal like tone modulation
> of an rf signal detected by the crystal. The amplification will only
> occur if the acoustic network is the equivalent of a lumped circuit.
> Resonance at 3kHz is not going to help listen to an ordinary voice
> signal.
I never said that it would. It's the combined effect of system resonances
at 1kHz (middle ear) and 3kH (ear canal) that together contribute to the
sensitivity of the human ear in the frequency range between 500Hz and 6kHz.
The point, which you apparently either missed or ignored, is that when it
comes to hearing sensitivity, an impedance mismatch can be and is an
advantage, and not, as you claim, something to be avoided.
> I never said that it would. It's the combined effect of system resonances
> at 1kHz (middle ear) and 3kH (ear canal) that together contribute to the
> sensitivity of the human ear in the frequency range between 500Hz and 6kHz.
> The point, which you apparently either missed or ignored, is that when it
> comes to hearing sensitivity, an impedance mismatch can be and is an
> advantage, and not, as you claim, something to be avoided.
Even if maximum pressure were wanted for specialize listening
situations, you would also want matching. You would want to match
impedance with the appropriate resonant structure's dissipation. Assume
a high Q resonator. For a series (electrical) resonator that would
typically mean using an acoustic matching transformer to a low impedance
(force/velocity) for the vibrating structure. For a parallel resonator
you would have to transform source impedance to a higher level.
One caveat is that a mechanical schematic diagram for a series resonance
looks somewhat like an electrical parallel resonant structure and vice
versa. Remember that the series resonance will have equal velocity for
the mass and the spring. In the diagram the mass will resemble the
symbol for a capacitor in parallel with a coil (inductor} which look
like a spring.
For whatever reason, you are still not listening. You can not talk about
building a high-sensitivity headphone when the headphone is to be used in
a passive circuit such as a crystal radio. This is so because the load
impedance of the headphone affects the terminal voltage of the circuit to
which it is connected and because the source impedance of the circuit to
which the headphone is connected affects the acoustic output of the
headphone. It is because of this interaction that the sensitivity of a
headphone is specified in terms of the pressure that it produces in
specified aoustic load (eg IEC-318 coupler) per applied volt when driven
from a low-impedance voltage source.
> Taken to the extreme, you're saying no headphone is any better than any
> other when
> connected to a crystal radio. Bullshit! I understand impedance matching, I
> understand
> loading the tank to an unusable Q. Go look at this 4 Megohm to 8 ohm audio
> transformer.
> http://www.crystal-radio.eu/entrafounit2.htm
> If you don't like my semantics get over it, everyone else understands I
> want maximum
> audio volume for the input power available.
> You are welcome to ignore this thread.
I am glad you have made a response. I thought mine was the only voice in
the wilderness. When I was a kid before I truly understood conservation
of energy and impedance matching, I thought that the higher the
impedance of headphones, the better. I did not realize that all
headphones of the same design would have essentially the same
efficiency. I did not understand why you would not want to get the
largest possible number of turns on the phones.
It took me a long time to understand why a 5-tube superheterodyne
receiver could be as sensitive as it was. Now, I realized that you tried
to match the source of rf power to the losses of the parallel tuned
intermediate frequency transformers. The loss of power to the grid of a
pentode was minuscule compared to that in the resonant circuit. Matching
gave the highest grid voltage to a tube amplifier that responded in
proportion to the applied voltage. Loss in the tuned circuit would have
to be great enough to allow adequate signal bandwidth.
The same principles apply to acoustic and mechanical devices.
I agree with your conclusion. The problem is that what you say I said
isn't what I said or even implied.
> I understand impedance matching, I understand loading the tank to
> an unusable Q.
That's a good start. .
> Go look at this 4 Megohm to 8 ohm audio transformer.
> http://www.crystal-radio.eu/entrafounit2.htm
Interesting but irrelevant.
> If you don't like my semantics get over it, everyone else understands
> I want maximum audio volume for the input power available.
Your semantics aren't the problem.
> You are welcome to ignore this thread.
> MikeK
I'm not going to ignore this thread because I'm curious about what
valuable information others may provide regarding the subject matter.
But I am going to ignore you.
>> Go look at this 4 Megohm to 8 ohm audio transformer.
>> http://www.crystal-radio.eu/entrafounit2.htm
>
> Interesting but irrelevant.
>
Not irrelevant if you want to connect a low impedance headphone to a
detector with a high output impedance.
At this point I don't know the output impedance of high Q tank driving
a diode that matches the tank, but I do think 200Kohm to 2.5megohm
are not unreasonable numbers*.
MikeK
I like factors greater than 10, almost easier the horseshoes to get close.
;-)
(it does give a range and tank Qs vary a lot in crystal radio building)
Yeah, the old electromagnetic headphones were high impedance;
typically two coils interacting with a light magnetic plate. My father
gave me a pair when I was young and they did work in crystal sets. I
don't recall what their impedance was, but I want to say 10kOhms.
What a great link amdx. Thanks. I hope to try to make my own diode one
day and have just started playing around with metal oxides in my
woodstove, simply by putting in a tin can top and cranking the heat.
Plenty of resistance and material that cracks off when bent, which can
sort of be scratched through to the metal- must be oxides. I don't
mind primitive and poor results when starting out... no diodes yet,
but possible resistors?
Good luck with the headphone. I believe you will get it, but yes, lots
and lots of fine windings. Some printer stepper motors have pretty
fine wire windings, and the wire is in rolls, so is easy to reuse.
It's not ultrafine, but it doesn't look like you'll need it with that
contraption. Looks like a good prototype design. Don't forget, DC
resistance and AC impedance are two different things.
- Tim
Agreed. The main difficulty in our understanding of the CR system is
that at this time no model including the transformation of the electrical
power derived from the radio wave into acoustical power radiated from a
speaker transducer has been advanced. The "Gallows" phone is an example of
effort to improve the conversion of electrical audio power (demodulated RF
power, no amplification) into sound one can hear. When speaking of
reproduced sound pressure level, the distance at which that is heard is not
yet in the model as far as I can see.
> It is because of this interaction that the sensitivity of a
> headphone is specified in terms of the pressure that it produces in
> specified acoustic load (eg IEC-318 coupler) per applied volt when driven
> from a low-impedance voltage source.
..... which is a human ear simulator... It's a start. The "distance" at
which the sound is heard is about an inch.
(My mnemonic for conversion of sound power data to SPL in dB is that the
sound pressure found one foot from a point source (small speaker) is
numerically equally to the sound power expressed in picowatts (10^-12 W)).
So the sound dB sound level created in that coupler's geometry is
numerically about 20 dB more than its sound power value... A +10dB sound
power level produced by the transducer under test would then produce a +30
dB SPL in that coupler.... (ad-hoc WAGs by me).
(FYI, a vacuum cleaner emits about +85 to +90 dBA sound power).
Ange
--------------------------------------------------------------------------------
"The big question here is: Do phones that produce 125 dB SPL when 1 mW is
applied to them . . . produce 35 dB SPL when 1 pW is applied ?
1 pW is 90 dB below 1 mW. I believe 35 dB SPL should be very comfortable
listening. 1 pW audio might be what a crystal radio would deliver from a
weak DX signal.
Much debate has transpired in the past over this issue, without any definite
conclusions. Anecdotally speaking: some Sennheiser "120 dB/mw" earbud-type
phones failed miserably on weak DX when compared to a good set of
sound-powereds."
MikeK
For either a circumaural headphone or an insert earphone, which is what
is required for this application, it's the enclosed air volume, not the
distance between the transducer and the ear that matters. This is so
because the acoustic impedance looking into the ear canal is stiffness
controlled (approximately 2cc equivalent air volume) in the frequency
range below about 1kHz. Consequently, up to about 1kHz, pressure at the
eardrum is proportional to the volume displacement of the diaphragm of
the transducer. Therefore, the best transducer for this applicaiton is
one that provides a the greatest volume displacement per applied volt,
that minimizes the amount of added air volume (relative to 2 cc) and that
has a critically-damped, lumped-element diaphragm resonance at about 3
kHz.
Dont even think about reusing enamelled copper wire
NT
If you want maximum sensitivity, the last thing you want is any
damping. Damping reduces output massively. Early moving iron horn
speakers had no damping, in band resonance, linearity issues etc, but
were still quite listenable, despite the quality. Or would that be
horribility.
NT
> On Nov 7, 10:38 pm, Answerman <answer...@home.com> wrote:
>> "Angelo Campanella" <a.campane...@att.net> wrote
>> innews:ib704f$s9g$1@news
Only if you are interested at output at specific single frequencies.
> Early moving iron horn
> speakers had no damping, in band resonance, linearity issues etc, but
> were still quite listenable, despite the quality. Or would that be
> horribility.
> NT
What do early moving iorn horn speakers that had no damping, in band
resonance and linearity issues have to do with earphones for use in a
crystal radio?
i didn't imagine my crystal radio headphones had flat frequency
response over the audible spectrum, but i don't think for an
application where the headphones are powered solely from an unpowered
device where the energy driving the headphones is what is scooped outa
da air, i don't think you want any resistive components that don't
come naturally with the wound coils or in the crystal diode (or the
headphone windings). perhaps there is a transformer for optimal
impedance matching and maximum power transfer, and there is some loss
in that transformer.
but you don't want any damping put in there. you deal with the
resistance you can't avoid (and, perhaps, make use of it to broaden
out some frequency response).
r b-j
no, I mean for audio speech & music use.
> > Early moving iron horn
> > speakers had no damping, in band resonance, linearity issues etc, but
> > were still quite listenable, despite the quality. Or would that be
> > horribility.
> > NT
>
> What do early moving iorn horn speakers that had no damping, in band
> resonance and linearity issues have to do with earphones for use in a
> crystal radio?
Point is they had the downsides of no damping, yet were good enough
for speech & music. Ie it is realistic to use such an approach to
maximise sensitivity for speech & music.
NT
Geeze, you've shot my whole approach. I must have about ten different
sizes or recycled enamel wire. I haven't just thought about it, I've
done it. It's true that some windings are better than others for reuse
though.
I have wound a few toroids, and hear that the level of detail does
descend down to whether the copper is close to the core, which for the
rectilinear cross section forms poses a problem for any heavy wire,
but certainly work hardened will be the worst. I've heard of guys
annealing the stuff, but I haven't gotten that far with it. Any tips
are welcome; often just a few short words go beyond any reference
material, particularly when recycling parts. Anyway, the DIY approach
will always be more enlightening; partially for the small sidetracks
and sensetivities that are learned along the way, and though we may
not get as much done the discoveries along the way are worthwhile. The
stepper motor wire that I have is on a nice diameter roll, straight
out of the motor.
For this project of the large diaphragm headphone I think there is an
interesting puzzle of selection of the winding core, and what it's
proper material and diameter should be. It is possible that winding,
say three, smaller cores could be superior to the single large core,
which could then also give some nice experimental feedback if there is
variation amongst the three cores. Particularly the ability to
(de)tune the diaphragm will come in here in terms of resonances and so
forth and placement of coils. This is a ton of dynamics to play with.
The old style headphones use two coils off center, on small
rectangular iron cores. I've been wanting to build a transducer for
some time and one day I'll get to it; probably from a coffee can, and
admittedly the first will just use one coil, where max gain is likely
at the fundamental frequency of the diaphragm, which then should be at
say 600 Hz or so. This would be an ideal CW transducer, and I don't
know why they are not on the market; tune it right up!
- Tim
Classical room acoustics (the earcup on head is a room for this
discussion) has it that regardless of the volume, the sound level in a room
is calculated directly from the sound power inserted and the absorption,
volume does not matter (!!). In decibels,
SPL = PWL-1*log(A) + 6 (A in square meters)
SPL = PWL-10*log(A) + 16.3 (A in square feet)
One picowatt is 0dB PWL
Assume the cup and ear to have a surface area of about 30 square
inches (0.2 square feet), with an absorption coefficient 50%, A = 0.1,
so the 10*log(A) is -10 dB.
This leaves us with 6 dB of SPL. At 1 or 2 kHz, this is ( for a good
ear, fellas) about at threshold...
Now, that's 1 picowatt of Sound Power...
Considering audio transducer conversion efficiency of 10%, the
electrical-audio-frequency power needs to be 10 picowatts of audio ac.
Figure a factor of ten for demodulation (in)efficiency of 10% ups
this level to 100 picowatts..
The RF power that engendered that, let's say ten times more, or 1,000
picowatts (one microwatt).
Therefore AM RF absorbed or captured by the crystal set and its antenna
must be more than one microwatt of RF energy.
> Therefore, the best transducer for this application is
> one that provides a the greatest volume displacement per applied volt,
> that minimizes the amount of added air volume (relative to 2 cc) and that
> has a critically-damped, lumped-element diaphragm resonance at about 3
> kHz.
I think we can assume that he earcup volume SPL is uniform throughout
for frequencies up to about 2 kHz.
The ear canal resonance will be as it is for a open ear.
Ange
> Considering audio transducer conversion efficiency of 10%, the
And there is the where I started! How do I increase that 10% to 20% or 30%?
The easy 1 to 3db is to eliminate the impedance transformer, so make the
speaker impedance high, with a tapped coil (for matching) on the speaker.
(I don't know if it is feasible to make a speaker coil with 500kohm
impedance)
Or a multi peizo speaker with series/parallel switching arrangement to alter
the impedance.
All this used in the gallows style diaphragm.
Thanks, MikeK
Proper enclosure design.
Good Luck!
Rich
No reason to go any further, because your numbers thus far don't comport
with reality. One of the more efficient headphones, the Senheiser HDA-280,
produces 117dB SPL at the eardrum at 1kHz for 1mW of electrical input, or
equivalently, 27dB SPL at the eardrum at 1kHz for 1pW electrical input. So,
without going into why, the fact of the matter is that your calculation
thus far is in error at least 20dB.
If I understand your analysis correctly, your bottom line is that 10pW of
electrical power is required to produce 6dB SPL at 1kHz at the eardrum.
If so, there is no reason to go any further, because your numbers thus
far don't comport with reality.
One of the more efficient headphones, the Senheiser HDA-280, produces
117dB SPL at the eardrum at 1kHz for 1mW of electrical input, or
equivalently, 37dB SPL at the eardrum at 1kHz for an electrical input of
10pW. So, without going into why, the fact of the matter is that your
calculation thus far is in error at least 30dB.
Remember, guys, these are all WAG's, just to get an analysis framework
up.
We can tweak as we go along.
> One of the more efficient headphones, the Sennheiser HDA-280, produces
> 117dB SPL at the eardrum at 1kHz for 1mW of electrical input, or
From one milliwatt to one picowatt is -90 dB . OK. = 27dB for 1 pw.
> equivalently, 37dB SPL at the eardrum at 1kHz for an electrical input of
> 10pW. So, without going into why, the fact of the matter is that your
> calculation thus far is in error at least 30dB.
Not too bad for a beginning system analysis, I say... This relates to
the audio sensitivity that can be achieved.
37 dB @ 1 kHz is nicely audible in a quiet room, so that the HDA-280 should
be useful to use with a crystal receiver. But what is the input impedance of
that headset? if it is 8 ohms or even 50 ohms, there will be a impedance
mismatch of as much as 30 dB....
Ange
I think you are missing the fundamental quandary,The major cause of
inefficiency of a speaker is poor impedance matching between the speaker
cone and the wave you are trying to propagate in the air. It is NOT the
mismatch between an amplifier and putting power into the voice coil or
diaphragm.
Is this a good assumption? Will the audio output SPL be linear with the
power input?
Is the answer the same for all transducer types?
Thanks, MikeK
--------------------------------------------------------------------------------------------------
THIs from my previous respoNSE
"The big question here is: Do phones that produce 125 dB SPL when 1 mW is
applied to them . . . produce 35 dB SPL when 1 pW is applied ?
1 pW is 90 dB below 1 mW. I believe 35 dB SPL should be very comfortable
listening. 1 pW audio might be what a crystal radio would deliver from a
weak DX signal.
Much debate has transpired in the past over this issue, without any definite
conclusions. Anecdotally speaking: some Sennheiser "120 dB/mw" earbud-type
phones failed miserably on weak DX when compared to a good set of
sound-powereds."
---------------------------------------------------------------------------------------------------
> Ange
The nominal impedance of the HDA-280 is 37 ohms. Because of the relatively
low impedance, it is not a headphone that I would be my first choice for
this application. I simply used it as an example of the efficiency that is
attainable using current technology in the design of a dynamic headphone.
Also, increasing the impedance of the headphone with a transformer is
problematic even if the transformer were ideal. The reason is because for
every N^2 increase in impedance that the transformer provides, there is a
1/N decrease in overall transformer plus transducer sensitivity when
measured in terms of eardrum pressure per volt applied to the transformer.
I don't know if there is an N that optimizes overall transformer plus
earphone sensitivity for a known earphone impedance and a desired load
impedance on the diode demodulator. Perhaps someone else can work out the
math.
> It's not that I'm missing it, you'll note I said;
> "The >easy< 1 to 3db is to eliminate the impedance transformer".
> The quandary is that I don't know how to do the mechanical matching.
> As far as the piezo, I think that is just a matter of picking the right
> piezo?
> (assuming a completed bender)
It is typically very difficult to get broadband matches when there is a
fundamental mismatch. That applies to antennas, multilayer thin-film
antireflection coatings, and speakers.
> What do early moving iorn horn speakers that had no damping, in band
> resonance and linearity issues have to do with earphones for use in a
> crystal radio?
Just everything, "Answerman". Oddly nobody here seems to have noticed
that the famed "gallows" headphone is nothing but a re-invention of
the standard old-time telephone earpiece. And I'd point out those
things are QUITE sensitive being used for crystal sets from the
beginning. The horns in question were ALSO able to work with crystal
sets though usually not very loud. The high sensitivity and high
impedance of such transducers made them very suitable. "Ear buds" in
spite of the sensitivity and not as good because they are low
impedance (require a matching transformer) and they are designed to
have wideband flat response. As Tesla demonstrated on many occasions
if you really want some action you need resonant devices. Little
damping as possible. You are all wandering in the wilderness here.
WOW! The first speaker working like that I ever saw was while cleaning
out the finished attic of an old house circa 1950. That speaker was built
exactly like that. I recall it to be a large cone like you show, maybe a
foot or more in diameter. The coil driver was just a big old relay
coil-looking sort of thing. The lever and its fulcrum looked just like a
relay armature; the cone driver just an extension rod on the other side of
the armature bearing-shaft. This is an easy way to get more leverage, a
mechanical impedance transformer if you will. That unit was obviously a
refugee from an ancient AM radio from some where in 1920's. The former owner
was a past Mayor of Wilkes-Barre, PA, Dan Hart... He was sai dto have had
some wild parties up in that finished garret during Prohibition days.
Anyway, the important factors are the steps made to mechanically
"impedance match" the electrical force available; a strong armature force
which can only go a short distance over coil - to the cone wagging in the
air presenting little reaction force, but should travel over a much longer
distance.
Go for it... build one..
You could use an old high impedance relay, mounted with coil axis
horizontal, on an inertial base (e.g. a brick) .
The mechanical transformer ratio is adjusted as the length of the
extension rod.
It was about an inch or so long as I recall...
The cone drive point has an axial pin glued into the apex tip of the
cone; use super-glue, I think.
Manila folder stock, painted with clear dope to stiffen it, may work as
the cone. I'm not sure of the cone base support, but I suspect that any
hanging arrangement will do; the axis of the cone is horizontal.
This could be a "Big-Gallows" speaker; completely impedance matched to
air.
Ange
According to the diagram, the acoustic load seen by the speaker cone is a
confined volume of air, the impedance of which is a purely reactive
compliance. In order to match that load impedance, the source impedance
would need to be a purely mass reactive, and the match would only exist
at a single frequency. For a broadband (300Hz-3000Hz) impedance match,
both the source and load impedances would need to be predominantly real
and approximately equal in magnitude. I don't see the latter requirment
being satisfied in the proposed design. What am I misssing?
Hi Ange,
This is basically the same arrangement found in some of the older headsets.
Although being the Gallows style it can be built big, so fat fingers can do
it,
also the size will allow many turns for impedance matching. The reason I
went with a manufactured speaker assembly is because maybe the solved
some of the matching problems that I'm clueless about. Although by sealing
the speaker and adding the tube, I've probably messed up all the good design
that the speaker had. I don't really like the fulcrum idea as drawn, but I
have
another idea (probably not new either) that I'll draw up later.
Re:
>This is an easy way to get more leverage, a mechanical impedance
>transformer if you will.
It would be neat to have an adjustable fulcrum to fine tune the
leverage, I don't know where the trade off is between cone
displacement and the power available to overcome the leverage.
Now to find/build an iron core with a gap as drawn.
Are the laminations thinner on audio transformers than on power
transformers?
Any tape wound cores used for audio? (might be easy to mill the gap)
Mike
If it isn't already apparent to everyone, you've spotlighted the two key
requirements which are high electroacoustic sensitivity and high
impedance so as to not load the detector. In order to achieve both
simultaneously, the use of an impedance matching transformer between the
detector and the earphone needs to be avoided. This is so because for
every increase N^2 increase in impedance, there is a 1/N decrease in
overall trasformer-earphone sensitivity. The ideal solution is to
increase the earphone impedance as much as possible while still
maintining the hightest possible sensitivity. Because of present mass-
market requirements, the focus of current earphone technology is on high
output and high efficiency, and not on ultra high impedance.
Consequently, current technology is not the place where on is going to
find an improved high sensitivity, ultra high impedance earphone that is
optimized for use with a crystal radio. So the only solutionare to
either build your own earphone from scratch or to modify an existing
balanced armature. The good news is that it's already been done.
http://www.crystalradio.us/headphones/index.htm
If the results can be believed, they are impressive. Specifically,
an AC Impedance at 1 kHz of 56000 Ohms (pair) and a threshold sensitivity
at 1 kHz of 0.00028 picoW (pair). For someone with young-adult normal
hearing, the latter translates approximately to 35dB SPL for 1pW
electrical input. So these mofified headphones have a sensitivity that
is comparable to that of the Sennheiser HDA-280, but with an impedance of
56000 ohms instead of 35 ohms. Consequently, no matching transformer is
required and the associated loss in overall transformer-earphone
sensitivity is avoided.
I'm not understanding this this.
Do you mean the voltage to the driver is divided by the turns ratio?
> The ideal solution is to
> increase the earphone impedance as much as possible while still
> maintining the hightest possible sensitivity. Because of present mass-
> market requirements, the focus of current earphone technology is on high
> output and high efficiency, and not on ultra high impedance.
> Consequently, current technology is not the place where on is going to
> find an improved high sensitivity, ultra high impedance earphone that is
> optimized for use with a crystal radio. So the only solutionare to
> either build your own earphone from scratch or to modify an existing
> balanced armature. The good news is that it's already been done.
> http://www.crystalradio.us/headphones/index.htm
> If the results can be believed, they are impressive. Specifically,
> an AC Impedance at 1 kHz of 56000 Ohms (pair) and a threshold sensitivity
> at 1 kHz of 0.00028 picoW (pair). For someone with young-adult normal
> hearing, the latter translates approximately to 35dB SPL for 1pW
> electrical input. So these modified headphones have a sensitivity that
> is comparable to that of the Sennheiser HDA-280, but with an impedance of
> 56000 ohms instead of 35 ohms. Consequently, no matching transformer is
> required and the associated loss in overall transformer-earphone
> sensitivity is avoided.
This bit of research is not worthy of the big promotion I gave you!
The key words on the page you reference are;
"Since his pair of phones has a 56k impedance at 1 kHz, a matching
transformer
can be eliminated for SOME crystal set applications."
This design is for a contest quality transducer, thus we can assume a high
quality crystal radio.
A good inductor can have Q of 900 at 1500khz or a loss resistance of 2.5
ohms (R1).
A capacitor capacitor can have a Q of 5000 at 1500khz or a loss of 0.45 ohms
(R2).
Using the formula Z = (2 x pi x F x L) ^2 / R1 +R2 the impedance at
resonance is
1.725 Megohms.
Now we install a diode (detector) that matches the 1.725 Mohms at this
signal level,
(the diode resistance changes with signal levels).
Now I don't know exactly where the impedance is at this point, but it is
much higher than
56,000 ohms. So, the 56,000 ohms headphones are not that good of a match to
a high
quality crystal radio.
So I'm back to asking for at least 1megohm transducer impedance.
Thanks for your interest, Mike
Thanks for that, I know of them, wouldn't know how to implement its use.
You have included it on our list of tools for possible use in the HSCRHP.
Thanks, MikeK
Been thinking about this some and it makes sense that the coil pair
would be part of the same magnetic circuit, so that for a magnetic
diaphragm the area between the two coils was conducting
(magnetically), with the loop passing down one coil core and out the
other one. This is the best means to a short magnetic path through the
light diaphragm, at the position of greatest motion. This should be
easily verified with an old pair of headphones, for the cores of the
two windings will be a U of iron. To ease the weight of the new high
efficiency diaphragm this could be a single transformer plate cut down
and attached to a lighter weight diaphragm. Those old phones had a
circular thin ferrous plate. Ahh... how much does the mass really
matter? If at resonance no heating takes place then the mass is a tank
circuit, and merely rings a bit. Yes, this does matter quite alot, for
the peak intensities of the signal will become subdued, and no doubt
these are critical portions in terms of intelligibility at the limits.
I believe that you want small cores and small magnetics for low level
signals, so that eddy losses are minimal. This style will not do well
with big signals, but that is OK, right? This is magnetic saturation
kicking in, and that should set a smallest dimension. This is also
consistent with selecting the smallest diaphragm magnetic conductor,
so that a light weight motion is possible. I believe that the low
power constraint actually leads to a small device; one that may fit in
the ear. Still, the photo is an excellent prototype platform to work
from. Be careful with these things! It would be a shame to blow your
ear's sensitivity with a bad design. Mine already are abused, and I am
more careful about loud noises now.
Jeeze, I've got to get to some magnetics this winter.
- Tim
> I've been wanting to build a transducer for
> some time and one day I'll get to it; probably from a coffee can, and
> admittedly the first will just use one coil, where max gain is likely
> at the fundamental frequency of the diaphragm, which then should be at
> say 600 Hz or so. This would be an ideal CW transducer, and I don't
> know why they are not on the market; tune it right up!
>
> - Tim
I'm still wondering how you're going to design a high quality crystal
radio!
> A good inductor can have Q of 900 at 1500khz or a loss resistance of 2.5
> ohms (R1).
> A capacitor capacitor can have a Q of 5000 at 1500khz or a loss of 0.45 ohms
> (R2).
> Using the formula Z = (2 x pi x F x L) ^2 / R1 +R2 the impedance at
> resonance is
> 1.725 Megohms.
> Now we install a diode (detector) that matches the 1.725 Mohms at this
> signal level,
> (the diode resistance changes with signal levels).
> Now I don't know exactly where the impedance is at this point, but it is
> much higher than
> 56,000 ohms. So, the 56,000 ohms headphones are not that good of a match to
> a high
> quality crystal radio.
> So I'm back to asking for at least 1megohm transducer impedance.
> Thanks for your interest, Mike
Transducer loading doesnt exactly help with q. Also the shape of the
sensitivity curve is all wrong. No agc, very low sensitivity, etc
NT
"Answerman" wrote in message